L2s Now Handle 93% of Ethereum Activity—But L1 Fee Revenue Dropped 90%. Who's Paying for Base Layer Security?

General
1

The numbers don’t lie, and they’re starting to worry me.

The L2 Migration Is Complete—And It’s Spectacular

We did it. Layer 2s now handle 93% of all Ethereum transaction activity. L2s are processing over 1.9 million transactions per day, completely eclipsing mainnet. Base leads with 10.5M transactions in 24 hours, Polygon PoS with 9.4M, Arbitrum One with 3.2M. The top 10 L2s alone handle over 31 million transactions daily.

From a scaling perspective, this is exactly what we wanted. EIP-4844 and blob transactions slashed L2 operating costs by 90-99%, bringing daily costs down to around $135,100 while transaction volumes surged from 500K in February 2024 to 2.1M+ by October 2025. Users get near-instant transactions for pennies instead of dollars. The rollup-centric roadmap is working.

But There’s a Problem: The Economics Don’t Add Up

Here’s what’s keeping me up at night: L2 fee payments to Ethereum L1 have collapsed over 90% year-over-year. We went from $113 million in 2024 to just $10 million in 2025. That’s a 91% drop in revenue flowing to L1 validators.

Think about what this means:

  • Ethereum L1 security comes from 1.1M validators with 30% of ETH staked
  • Validators secure the network because of economic incentives: transaction fees + MEV + staking rewards
  • When 93% of activity moves off-chain, validators lose 93% of the fee revenue that funded those incentives
  • Meanwhile, L2 sequencers are capturing approximately $161M per year in sequencer fees—revenue that never touches L1 validators

We’ve successfully scaled Ethereum by moving activity to L2s. But in doing so, we’ve starved L1 of the fee revenue it needs to maintain security.

The Settlement Layer Paradox

Blob fees operate in a separate market, typically ranging from 1-10 wei compared to calldata’s 30-50 gwei. Blobs are currently running at about 30% capacity—which means even if usage doubled, we’re not generating meaningful L1 revenue from L2 settlement.

The Ethereum Foundation says L1 is becoming a “settlement layer”—and that’s fine in theory. But here’s the question: if L1 becomes infrastructure that nobody directly uses, who pays to secure it?

Some will argue that staking yields of 3-5% APY are sustainable even with lower fees. But that math assumes:

  1. ETH price appreciation offsets reduced fee income
  2. We don’t need as much security for a settlement layer as for an execution layer
  3. MEV will continue to provide meaningful validator revenue

I’m not convinced all three assumptions hold, especially in a bear market.

Potential Solutions—All With Trade-offs

Option 1: Based Rollups
If L2 sequencer revenue (~$161M/year) shifted to L1 validators via based rollups, it would significantly boost ETH staking yield. But this requires L2s to give up sequencer control and profitable businesses. How many L2s will voluntarily do this?

Option 2: Security Rent or Protocol Fees
Some EIP proposals suggest introducing a minimum execution base fee on Ethereum to restore sustainable burn and validator revenue. Alternatively, L2s could pay “security rent” proportional to the value they’re securing. But if we raise L2 costs, we defeat the purpose of scaling.

Option 3: Accept Lower Security Budget
Maybe L1 doesn’t need huge security budget if it’s just settling L2 transactions? Fewer validators, smaller rewards, but adequate for the role. The risk: if you reduce security budget, you make the network more vulnerable to attacks, especially if L2s collectively hold $40B+ TVL.

Option 4: Capacity Expansion (Futureproofing)
The Fusaka upgrade plans to increase blob capacity from current levels to eventually 64 blobs per block—a 40x throughput increase. PeerDAS (EIP-7594) will optimize data availability sampling. But this assumes demand will come, and if blobs are only 30% full now, why build 40x more capacity? And more capacity at lower prices means even less revenue per transaction.

My Question to the Community

As someone who’s spent the last 6 years building L2 infrastructure, I genuinely don’t know the answer here. The rollup-centric roadmap is technically sound—L2s are faster, cheaper, and more accessible than L1 ever was. But I haven’t seen a clear answer to the economic sustainability question.

Who should pay for L1 security when L1 becomes infrastructure nobody directly uses?

Should L2s pay security rent based on TVL secured? Should we mandate based rollups to redirect sequencer revenue? Should we accept that lower L1 activity means lower security budget (and hope it’s sufficient)? Or is there a fourth option I’m missing?

The data shows we’ve solved the scaling problem. Now we need to solve the sustainability problem before it becomes a crisis.

2

Lisa raises the right concerns, but I want to offer some perspective from someone who’s been contributing to Ethereum core development: the settlement layer model is not a bug, it’s the feature we’ve been building toward for years.

The Rollup-Centric Roadmap Was Always the Plan

When Vitalik outlined the rollup-centric roadmap back in 2020, the explicit vision was that L1 would become primarily a settlement and data availability layer, not an execution layer. The massive scaling gains (10,000+ TPS) were never going to happen on L1—they were always going to come from rollups using L1 for security and consensus.

So the fact that 93% of activity has migrated to L2s isn’t a failure of economics—it’s the successful execution of the architectural vision. L1 was never meant to compete with L2s for transaction throughput.

Security Budget: Do We Need as Much?

Here’s the part that might be controversial: I don’t think L1 needs the same security budget as a settlement layer that it needed as an execution layer.

When L1 was processing all transactions, attack surfaces were everywhere: smart contract exploits, MEV manipulation, front-running, sandwich attacks, protocol-level vulnerabilities across thousands of dApps. The security budget needed to be massive because the attack surface was massive.

As a settlement layer, L1’s attack surface is much narrower:

  • Validating L2 state commitments and fraud/validity proofs
  • Ensuring data availability for rollup transactions
  • Maintaining consensus on the canonical chain
  • Protecting against 51% attacks and finality reversion

This is still critically important, but it’s a different threat model. You don’t need the same economic security to validate proofs that you need to execute arbitrary smart contracts.

Based Rollups: The Missing Revenue Bridge

Lisa mentioned based rollups briefly, but I think this is actually the most promising path forward. For those unfamiliar: based rollups are rollups where L1 validators sequence L2 transactions instead of separate L2 sequencers.

Right now, L2 sequencers extract about $161M/year in revenue that could flow to L1 validators. If major L2s adopted based sequencing:

  • L1 validators capture L2 sequencer fees, boosting staking yields
  • L2s inherit L1’s decentralization and liveness guarantees
  • No need for complex “security rent” mechanisms
  • Economic incentives naturally align L1 and L2

The challenge is convincing L2s to give up profitable sequencing businesses. But as competition intensifies and users demand more decentralization, I think we’ll see more based rollups emerge.

Protocol-Level Revenue Mechanisms

There are also EIP proposals in discussion (though not yet formal) to introduce a protocol-level minimum execution base fee that would establish a floor on L1 revenue regardless of blob utilization. This would ensure validator rewards don’t decay to zero even if blob fees remain cheap.

Think of it like a “security rent” but built into the protocol rather than negotiated per-L2. L2s posting to Ethereum would pay a minimum fee that scales with activity, ensuring L1 maintains economic security proportional to the value it’s securing.

The Real Question: Transition Management

I think Lisa’s concern is less about the end state and more about how we manage the transition. Right now we’re in an awkward middle phase:

  • L1 is no longer capturing execution fees (moved to L2)
  • But L1 hasn’t yet captured L2 settlement fees at scale (blobs too cheap, only 30% utilized)
  • And based rollups haven’t achieved meaningful adoption yet

This transition period could last 2-3 years, and during that time, validator economics might be squeezed. I think that’s a legitimate concern that needs protocol-level solutions, not just “wait for the market to figure it out.”

But long-term? I’m optimistic. Settlement layers for $40B+ in L2 TVL will need robust security, and the market will pay for it—whether through blob fees, protocol minimums, or based rollup sequencing revenue.

The architecture is sound. We just need to make sure the economic model catches up to the technical reality.